📝 SummarXiv

Abstract

We present a CMB temperature power spectrum measurement at large angular scales from WMAP and Planck maps that were cleaned of foregrounds using a template-based approach described in a companion paper. We recover essentially the full-sky CMB information at multipoles $\ell<30$ with only 1% of pixels near the Galactic plane masked and no inpainting. Notable features continue to appear: (a) a low quadrupole power compared to the best-fit Planck 2018 $\Lambda$CDM spectrum at $2.2\sigma$, (b) a dip in the range $20 \le \ell \le 27$, and (c) an overall $\ell<30$ power level low of the $\Lambda$CDM prediction derived from higher multipole moments. Given the different methodology from previous analyses and the nearly full-sky solution presented here, these features do not plausibly arise from foreground contamination, systematic errors, masking, or mode-mixing. Our overall $\ell<30$ amplitude constraint is consistent with published WMAP (77% sky fraction) and Planck (86%) results at $1.2\sigma$ and $0.6\sigma$, respectively, accounting for the improvement in statistical precision. We present a new $\ell<30$ likelihood for use with the Cobaya package. Parameter constraints from joint fits with the higher-multipole Planck data are consistent with the published Planck results, for example we find $H_0=67.35\pm0.54$ km s$^{-1}$ Mpc$^{-1}$ in a joint $\Lambda$CDM fit.